CHAPTER 82 PREDICTED FLUSHING TIMES AND POLLUTION DISTRIBUTION IN THE COLUMBIA RIVER ESTUARY Victor T. Neal Assistant Professor Department of Oceanography Oregon State University, Corvallis, Oregon ABSTRACT The observed salinity distribution in the Columbia River Estuary is used to predict the flushing times for various river discharge rates, varying tidal ranges, and varying salinity intrusions. Both the modified tidal prism method and the fraction of fresh water method are used, and the results are compared. The latter method predicts shorter flushing times, while both methods vary m a similar manner with changing river discharge and changing salinity intrusion. Both methods predict a rela- tively short flushing time for the estuary. The observed salinity distributions under varying conditions are also used to predict the distribution of conservative and non-conservative pollutants. In this case the fraction of fresh water method and the dif- fusion equation are used. In general, the fraction of fresh water method predicts higher concentrations. "When the lower estuary is divided into two channels, the two methods give quite different results. The diffusion equation method predicts a peak concentration upstream from the outfall rather than at the outfall, when the outfall location is arbitrarily placed at certain locations. INTRODUCTION THE RIVER The largest river on the Pacific Coast of North America, the Columbia, begins in Columbia Lake, British Columbia, Canada, and flows some 1, 200 miles before discharging into the Pacific Ocean near Astoria, Oregon. The mam tributaries to the system are the Okanogan, Kootenai, Clark Fork, Snake, and Willamette Rivers. The drainage area of this system amounts to nearly 259, 000 square miles, most of it between the Rocky Mountains and the Cascade Mountains. A large portion of this rugged area is made up of lava flows and is known as the Columbia Plateau. The Columbia is the most important inland waterway m the Pacific Northwest. The river is navigable from its mouth to Vancouver, Washington, and a good volume of ocean traffic continues up the Wil- lamette River to Portland, Oregon. Barge traffic now extends upstream as far as Pasco, Washington, by means of lock systems in the many large dams that have been built on the river. 1463 1464 COASTAL ENGINEERING THE ESTUARY The width of the river changes from less than one mile at a distance of 50 miles upstream to approximately nine miles at a distance of 20 miles from the mouth (Figure 1). This section has many winding chan- nels which are separated by low islands, bars and shoals. The depths vary from shoal areas to natural depths of over 100 feet below mean lower low water. The ship channel is presently maintained at a depth of 3 5 feet. The section from just inside the mouth to the area opposite Tongue Point consists of two mam channels (Figure 2). The channel on the Oregon side is the dredged ship channel; the channel on the Washington side is not dredged at present. The channels, although not completely distinct, are separated by several low islands, bars and shoals. The effect of the two channel system on circulation is important since the north channel carries the greater flood current, but the weaker ebb current. The tide is the mixed semi-diurnal type typical of the west coast of North America. The mean range at the mouth is 6. 5 feet while the range from mean lower low water to mean higher high water is 8. 5 feet. The tide travels up the estuary like a progressive wave and may be noticed upstream as far as 140 miles during low river stages. Fresh water flow through the estuary varies seasonally. Even though the entire region is characterized by heavier precipitation m the winter than in the summer, the peak flows usually occur during the months of May through July due to melting snows in the high mountains. Low flows usually occur during the months of September through March, however, freshets may occur during this time. The maximum discharge of record, during the flood of June, 1894, was estimated at 1, 200, 000 cfs, while the extreme low river flow has been estimated at 65, 000 cfs (Hickson and Rodolf, 1951). The freshet flows average nearly 660,000 cfs. This wide range of river flow is expected to be controlled by dams which will hold the maximum flow to 600, 000 cfs and increase the minimum to 150, 000 cfs (Lockett, 1963). Salinity intrusion varies with the stage of tide and with the river flow. During low river flow and lower low water the salinity intrusion is less than 15 nautical miles but higher high water may increase this to nearly 20 nautical miles. "When the river flow is high the intrusion may vary between a maximum at higher high water of 13 nautical miles and a mini- mum at low tide stages of less than five nautical miles. Salinity data from many sources (O'Brien, 1952, Burt and McAlister, 1959, Corps of Engineers Report, I960) were compiled, compared, and used in classifying the estuary and predicting flushing times and pollution distribution. The classification of the estuary is quite complex since the degree of mixing varies from one part of the estuary to another and varies FLUSHING TIMES AND POLLUTION 1465 with tidal stage and river flow. During high river flow and high tide the estuary is partially mixed at the mouth, changing gradually to a strati- fied system upstream. As the tide ebbs, the estuary becomes stratified nearly to the mouth. Low river discharge and high tide produce a well- mixed situation near the mouth, gradually changing to a partially mixed system upstream. As the tide ebbs, the estuary becomes partially mixed as far downstream as the mouth. Small horizontal salinity gradients are found at the entrance during periods of maximum salinity intrusion. The salinity on the south side becomes higher than that on the north. However, the difference is not more than three parts per thousand. The Coriolis effect apparently does not control the water movement inside the mouth of the estuary since the greater flood velocities are found on the north side and the greater ebb velocities on the south side. This effect is probably pro- duced by the orientation of the two channels in the estuary (Figure 2). PREDICTION OF FLUSHING TIMES METHODS The modified tidal prism method. This method, presented by Ketchum (1951), predicts the flushing time for an estuary when the entire pollution load is brought into the estuary by the river. There- fore, the amount of river water within a given segment is an indicator of the accumulation of pollutants m that segment. The estuary is seg- mented according to the scheme which follows. The inner end is taken as that section above which the volume required to raise the level of the water from low to high water is equal to the volume contributed by the river during a tidal cycle. The water above this section is, therefore, entirely from the river flow and is fresh water. Each segment down the estuary is defined so that the high tide volume in the landward one is equal to the low tide volume m the adjacent seaward one. It is assumed that the water within a segment is completely mixed at high tide. Therefore, the proportion of polluted water removed on the ebb tide (the exchange ratio rn for that segment) is equal to the local mtertidal volume divided by the high tide volume of the segment, or ex- pressed in symbols, P r = (1) n P n +V n The river water found m each segment is a mixture of river water accumulated during several tidal cycles. If it is assumed that the river flow is constant during the period m question, then the total volume of river water, En, accumulated within any segment of the estuary at high tide can be found by the equation E n = -9-r (2)v ' n where Q represents the volume of river flow per tidal cycle. 1466 COASTAL ENGINEERING The average number of tidal cycles needed for the river water to move through a given segment, or m other words, the flushing time for that segment, is equal to 1/r since l/r„n = TT^ • (3) Then the flushing time for the entire estuary is the sum of the flushing times for the individual segments. This method can be modified to include stratified estuaries. How- ever, such a modification produces an increased flushing rate, and was not used in the prediction. The fraction of fresh water method. The flushing time can be predicted by using the known distribution of salinity or fresh water m the estuary. The fraction of fresh water, f, m this method is deter- mined from the equation (S - S ) S 6 f = s (4) s m which S is the salinity of the sea water entering the estuary and S is the salinity of the water in a given segment of the estuary. When the fraction of fresh water has been determined, the volume of fresh water can be determined. The flushing time is then given by the time necessary for the river flow to supply the fresh water volume. If this is done for each segment of the estuary, the flushing time for the estuary is the sum of flushing times for the segments. RESULTS The modified tidal prism method. In order to use this method the cross-sectional areas, low tide volumes, and high tide volumes were measured or computed from the U. S. Coast and Geodetic Survey Charts No. 6151 and No. 6152.